Subscribe to get new posts from mleg.tech in your reader of choice.
I'm running Jev, TypeSafe AI's textless System-1 model, as the judge in my eval harnesses to see if evals can run on every commit instead of once a night.
Your agent hallucinates a contract clause. Sends it to a customer. Then, an hour later, a dashboard somewhere turns red.
Uptime and latency can't tell you if an agent's output is correct. Why AI observability needs a different kind of tooling than the APM stack you already have.
Bedrock Agents is now Classic, closed to new customers, and in maintenance mode — what AgentCore actually changes, and what it means if you're already running agents in production.
Reducer bugs, state bloat, and the deep-merge fix that took a full day to trace — lessons from running multi-agent workflows on AWS Bedrock at scale.
Why fixed-size chunking quietly tanks retrieval quality, and what content-hash dedup buys you at scale.
<?xml version="1.0" encoding="utf-8"?><feed xmlns="http://www.w3.org/2005/Atom"><title>mleg.tech — Michael Legemah</title><subtitle>Notes on building AI systems that ship.</subtitle><link href="https://mleg.tech/feed.xml" rel="self"/><link href="https://mleg.tech/"/><id>https://mleg.tech/</id><updated>2026-09-23T09:00:00.000Z</updated><author><name>Michael Legemah</name></author><entry><title>I swapped my eval judge for a System-1 model. Here is what I'm testing.</title><link href="https://mleg.tech/blog/swapping-my-eval-judge-for-a-system-1-model"/><id>https://mleg.tech/blog/swapping-my-eval-judge-for-a-system-1-model</id><published>2026-09-23T09:00:00.000Z</published><updated>2026-09-23T09:00:00.000Z</updated><summary>I'm running Jev, TypeSafe AI's textless System-1 model, as the judge in my eval harnesses to see if evals can run on every commit instead of once a night.</summary><category term="Eval & testing"/></entry><!-- 5 more entries --></feed>